Recursive Superintelligence, an AI research startup co-founded by former Salesforce chief scientist and You.com founder Richard Socher, emerged from stealth with a large funding round aimed at building AI systems that can improve themselves without human intervention.
Funding and valuation
<cite index="2-2,2-3">The company came out of stealth having raised over $650 million at a $4.65 billion valuation, in a round led by GV, Google's VC arm, and Greycroft, with participation from chip makers Nvidia and AMD.</cite> <cite index="3-18">The round was described as heavily oversubscribed.</cite> The editorial framing of this story referenced Andreessen Horowitz and Founders Fund as backers, but primary reporting identifies GV, Greycroft, Nvidia and AMD as the disclosed investors.
Team
<cite index="3-10">Socher leads the company alongside seven co-founders: Yuandong Tian, formerly a research scientist director at Meta's Fundamental AI Research lab (FAIR), where he led work on reinforcement learning, Large Language Model (LLM) reasoning, and AI-guided optimisation; Tim Rocktäschel, a professor of AI at University College London and former principal scientist at Google DeepMind; Alexey Dosovitskiy, one of the authors of the Vision Transformer (ViT), the 2020 paper that reshaped computer vision research; Josh Tobin, formerly of OpenAI; Caiming Xiong; Tim Shi; and Jeff Clune.</cite> <cite index="3-11">Peter Norvig, co-author of *Artificial Intelligence: A Modern Approach*, the standard university textbook in the field, serves as an adviser.</cite>
<cite index="4-11">The company currently operates from offices in San Francisco and London and has expanded to more than 25 researchers and engineers.</cite>
Technical thesis
<cite index="10-8">Recursive Superintelligence's central bet is that the fastest path to artificial general intelligence — and ultimately superintelligence — runs through AI systems capable of improving themselves by analyzing their own performance, without human intervention at each iteration.</cite> In a public launch statement, the company wrote that <cite index="10-1">"the fastest path to superintelligence will be realized by AI that recursively improves itself, and does so via open-ended algorithms that drive endless innovation,"</cite> adding that it will initially focus on AI that improves AI before extending the approach to other scientific disciplines.
Socher has emphasised that the company's distinguishing methodology is "open-endedness." <cite index="6-7">"Our unique approach is to use open-endedness to get to recursive self-improvement, which no one has yet achieved,"</cite> he said in a post-launch interview, drawing analogies to evolutionary biology.
Roadmap
<cite index="3-13,3-14,3-15,3-16">Recursive Superintelligence has outlined a staged roadmap. The first step, according to company materials, is to train a system with the capabilities of "50,000 doctors" to automate AI scientific research itself. From there, the company plans to run what it calls a "Level 1" autonomous training system, with a public launch targeted for mid-2026. The funding will be used in part to secure the large-scale compute infrastructure required to run these experiments.</cite>
Competitive context
<cite index="3-22,3-23,3-24">What distinguishes Recursive Superintelligence from other major efforts is that none of the leading laboratories has organised an entire company around recursive self-improvement as its core commercial thesis. OpenAI, Anthropic, and Google DeepMind all use AI to assist their research workflows, but their businesses are built around selling models and API access. Recursive is betting that the self-improvement loop itself is the product.</cite> Other frontier-research peers include Safe Superintelligence, which is pursuing a safety-first approach to superintelligence research.
<cite index="10-11">The company has not disclosed detailed plans for how it will ensure the safety of systems designed to modify their own architecture without human oversight — a question that, given the stated ambitions, will attract considerable scrutiny as the company's work becomes more visible.</cite>